After posting yesterday’s explanation of how drunk driving isn’t as dangerous as you might think from reading MADD propaganda, I began to worry that people might get the wrong impression. Let me clear this up: Drunk driving is very dangerous.
Keep in mind, that we’re only talking about fatal drunk-driving. There are going to be a lot more non-fatal accidents that leave people injured but alive. Mark Bennett has a really good post on his blog that takes my numbers and estimates the non-fatal injury rate from drunk driving. It’s a lot of carnage.
Rare events like drunk driving deaths involve numbers that are far enough outside normal human experience that our intuition sometimes fails us, so it’s hard to tell if a 1 in 18,000 chance of death is a lot unless you compare it to something else.
According to this page of statistics, your chance of dying as an occupant of a car, driver or passenger, over an entire year is also about 1/18000. This suggests that driving drunk compresses an entire year’s driving risk into a single trip. If you drive once a day, that implies a relative risk ratio of 365-to-1.
On the other hand, I found this NHTSA document which suggests only a 10-to-1 relative risk of a fatal accident for a BAC around the legal limit of 0.08.
I’m not entirely sure how to reconcile these numbers, but the NHTSA report says that the relative risk doubles to 20 for a BAC of 0.15, and one of the referenced studies implies much higher relative risks for males under 21 as BAC rises:
The relative risk of receiving a fatal injury in a single vehicle crash increases steadily with
increasing driver BAC for both males and females in every age group with one exception… Among all male and female drivers, except those in the 16-20 group, the relative risk of
receiving a fatal injury is lower for drivers with a positive BAC under 0.02% than for drivers with 0.0% BAC. Remarkably, however, for the 16-20 age group, the comparable relative risk was substantially increased even at this low positive BAC, by 55% among males, and by 35% among females. Looking at relative risk across the six age and gender groups, we find that at a BAC of 0.035%, it was elevated by a factor between 2.6 and 4.6, at a BAC of 0.065%, by a factor between 5.8 and 17.3, at a BAC of 0.09%, by a factor between 11.4 and 52, at a BAC of 0.125%, by a factor between 29.3 and 240.9, and at a BAC of 0.220%, by a factor between 382 and 15,560.
This is getting a little beyond my ability to think about statistics without looking things up in a textbook, but I think these numbers imply that although even mildly drunk drivers are dangerous, most of the real danger comes from drivers with a BAC way over the legal limit.
(By the way, note that for drivers over 21, the statistics appear to indicate that a BAC of 0.02 is less dangerous than a totally-sober BAC of 0.00. The NHTSA report suggests that this is an artifact of the study methodology and statistics.)
Mark Bennett ends his post with this:
Like most all societal problems, if it’s not solved by us rational people it’ll be solved by the kooks.
It seems to be an emerging theme of Windypundit to emphasize that serious problems require careful and honest thought. So, Amen to that sentiment, other Mark.